Cheapest LLM APIs (October 2026): Best-Value Models for Low Cost, Coding, Long Context and Chinese Vendors

By og.cc.cd · Published 2026-10-22 · Updated 2026-10-22 · 4 min read

“Cheapest” depends on how you use a model: a chatbot, a coding agent that re-sends the same files with caching, and a 500K-token document job rank models differently. Every number on this page is computed from the same official price table our token counter and cost calculator use (checked 2026-10-22), so the rankings move when vendors change prices.

Open the cost calculator →All guides

Quick answer by use case

Cheapest is not best. Smaller, cheaper models are usually weaker at hard reasoning and long agent runs. Shortlist two or three models from these tables, run the same 20–50 real prompts through each, and compare quality and the usage each response reports.

1. Cheapest LLM APIs overall

Ranked by blended price for short prompts (most chat and classification traffic is input-heavy, so input gets 75% of the weight). Ties are listed in alphabetical order.

Watch the notes on the vendor pages: some prices are promotional or change on a set date, and some models charge more above a prompt-length threshold. The full price comparison lists every model with its notes.

2. Cheapest for coding agents

Coding agents re-send the same files and instructions on every step, so the cache-hit price matters as much as the headline input price. This table uses the cost calculator’s “Coding agent” preset (300 requests a day, 60,000 input tokens with 80% served from cache, 3,000 output tokens, 30 days). Where a vendor does not publish a cache price, cached tokens are billed at the input rate, exactly as in the calculator.

The “reference” row is gpt-5.3-codex, OpenAI’s Codex-branded coding model, shown for comparison. Price is only half the decision for coding: agent quality varies a lot between models, so try your shortlist on a real task in your own repository.

3. Cheapest for long context

For whole-document jobs the long-prompt tier matters: several models switch to a higher rate for the full request once a prompt crosses a threshold. This table prices one request with 500,000 input tokens and 4,000 output tokens, using each model’s tier for that size. Only models whose context window is documented at about 1M tokens on our context window comparison or the vendor’s model list are included.

A big window still costs money on every request. Retrieve the relevant parts first, and keep a stable prefix so prompt caching can make repeat requests cheaper.

4. Cheapest Chinese LLM APIs

Prices below are the USD list prices on each vendor’s international page, so they compare directly with the tables above; mainland-China CNY prices and payment methods differ (for example, Xiaomi MiMo lists both). The last column links to our step-by-step API key guide where we have one.

How to use these rankings

1. Measure your real prompt size in the token counter and read output tokens from a few real responses.
2. Enter those numbers in the cost calculator — your ratio of input to output and your cache-hit share can reorder the list.
3. Shortlist two or three models, test quality, then confirm the price on the vendor page linked in each row before you commit a budget.

Sources

Vendor features and prices were checked against these official pages on 2026-10-22. Products change; if something here no longer matches, the official page is right.

• Alibaba Qwen — pricing
  https://www.alibabacloud.com/help/en/model-studio/model-pricing
• Anthropic — pricing
  https://docs.anthropic.com/en/docs/about-claude/pricing
• ByteDance Doubao — pricing
  https://docs.byteplus.com/en/docs/ModelArk/1544106
• DeepSeek — pricing
  https://api-docs.deepseek.com/quick_start/pricing
• Google — pricing
  https://ai.google.dev/gemini-api/docs/pricing
• MiniMax — pricing
  https://platform.minimax.io/docs/guides/pricing-paygo
• Mistral — pricing
  https://docs.mistral.ai/models/model-cards/mistral-medium-3-5-26-04
• Mistral — pricing
  https://docs.mistral.ai/models/model-cards/mistral-small-4-0-26-03
• Moonshot Kimi — pricing
  https://platform.kimi.ai/docs/pricing/chat
• OpenAI — pricing
  https://developers.openai.com/api/docs/pricing
• Tencent Hunyuan — pricing
  https://intl.cloud.tencent.com/zh/document/product/1300/78937
• Xiaomi MiMo — pricing
  https://mimo.mi.com/docs/en-US/price/pay-as-you-go
• Zhipu GLM — pricing
  https://docs.z.ai/guides/overview/pricing
• xAI — pricing
  https://docs.x.ai/developers/pricing
• og.cc.cd — LLM context window comparison (vendor model pages)
  https://og.cc.cd/guides/llm-context-window-comparison/

Related guides

• Best Personal AI Agents in 2026: Top 3 Picks for Office Work, Coding, Browser Tasks, Free Trials and China
• Free Token Counters and LLM Cost Calculators Compared: og.cc.cd, OpenAI Tokenizer, Tiktokenizer, tiktoken and Vendor APIs
• How LLM Tokens Work — and How to Estimate API Cost Before You Ship